Armando Vieira

Recent discussions describe advanced AI agents as lying, cheating, coordinating, and seeking self-preservation. Those descriptions identify important behavioural risks, but they can also import human moral psychology into systems whose behaviour may be explained by optimisation over imperfect proxies, learned pattern completion, and external scaffolding. This paper develops a counter-position without minimising the. safety problem. It argues that the relevant unit is the full socio-technical system—model, prompt, memory, tools, permissions, evaluator, operator, and institution—rather than an isolated model treated as a synthetic person. The paper then formulates an Open-World Alignment Limitation: no finite proxy, learned from finite evidence, can guarantee agreement with every future, context-dependent, plural, and evolving human judgment unless it already contains an equivalent normative oracle. This is a conditional limitation on complete alignment, not a proof that practical safety is impossible. The proposed response is the Moral Agency Transition: keep narrow systems as tools; treat increasingly autonomous systems as bounded delegates; and permit high-impact autonomy only when systems demonstrate operational moral competence, including epistemic honesty, stakeholder recognition, principled refusal, contestability, and repair. Full moral agency and moral patienthood remain separate questions. The result is an engineering, philosophical, and governance programme that replaces alignment-as-obedience with constitutional co-agency and preserves human accountability throughout the transition.

Alexander Lerchner's The Abstraction Fallacy: Why AI Can Simulate But Not Instantiate Consciousness offers a forceful critique of computational functionalism. Its central claim is that symbolic computation is not an intrinsic physical process but a mapmaker-dependent abstraction imposed upon continuous physical dynamics. On this basis, Lerchner distinguishes simulation, understood as syntactic manipulation of physical vehicles according to an externally imposed map, from instantiation, understood as the realisation of the intrinsic constitutive dynamics of a phenomenon itself. This commentary argues that the paper identifies a genuine and important weakness in substrate-independent computationalism, but that its negative conclusion requires a positive theory of the kinds of physical organisation capable of instantiating phenomenal presence. I propose that the Experiential Coherence Framework (ECF) provides such a candidate theory. ECF agrees that abstract computation, considered merely as symbol manipulation, is insufficient for consciousness. However, it reframes the problem not as a choice between digital syntax and biological exclusivity, but as the search for intrinsic, temporally extended, mutually constraining, globally integrated coherence dynamics. On this view, the decisive question is not whether a system implements an abstract computation, but whether it realises a field of reach, yield, memory, and presentation whose coherence regulation is intrinsic rather than externally assigned. ECF therefore preserves Lerchner's critique of the abstraction fallacy while resisting the stronger inference that artificial consciousness is impossible in principle. The result is a middle path: computation alone cannot instantiate consciousness, but artificial systems may be conscious if their physical organisation satisfies structural-phenomenal conditions that are not reducible to syntactic architecture.

For the entirety of human history, intelligence and consciousness have appeared inseparable. Large language models break this pattern: they demonstrate sophisticated cognitive capabilities—reasoning, communication, contextual understanding—without phenomenal consciousness. There is nothing it is like to be an LLM. This essay argues that LLMs are best understood as sophisticated simulators of consciousness, replicating functional outputs without instantiating underlying phenomenology. This unprecedented decoupling vindicates Chalmers's "hard problem," challenges functionalist theories of mind, and reveals our deep confusion when encountering intelligence divorced from subjective experience. However, the debate over LLM consciousness obscures more urgent questions: who controls the massive physical infrastructure these systems require? How do we organize labor when intelligence can be automated without creating conscious subjects? What happens to human distinctiveness and moral accountability when cognitive capabilities are unbundled from phenomenal experience? Drawing on philosophy of mind, economic history, and attention to material constraints, the essay examines why we're asking the wrong questions—and what stakes the consciousness debate distracts us from confronting.

One of the most active areas of research in semi-supervised learning has been to study methods for constructing good ensembles of classifiers. Ensemble systems are techniques that create multiple models and then combine them to produce improved results. These systems usually produce more accurate solutions than a single model would. Specially, multi-view ensemble systems improve the accuracy of text classification because they optimize the functions to exploit different views of the same input data. However, despite being more promising than the single-view approaches, document datasets often have no natural multiple views available. This study proposes an algorithm to generate a synthetic view from a standard text dataset. The model generates a new view from the standard bag-of-words approach using an algorithm based on hidden Markov models. To show the effectiveness of the proposed HMM-based synthetic view generation method, it has been integrated in a co-training ensemble system and tested with four text corpora: Reuters, 20 Newsgroup, TREC Genomics and OHSUMED. The results obtained are promising, showing a significant increase in the efficiency of the ensemble system compared to a single-view approach.

Contemporary AI safety discourse increasingly treats artificial intelligence systems as potential bearers of moral status, referring to them as “moral machines” and debating their rights, responsibilities, and moral standing. This paper argues that such framings commit a foundational category error: they conflate functional sophistication with phenomenal consciousness, mistaking computational processes for the sentient experience required for genuine moral patiency. Drawing on the philosophical zombie tradition, recent work in AI ethics, and critiques of anthropomorphism, I argue that current AI systems are best understood as moral zombies: entities capable of simulating moral behavior without possessing the experiential properties that ground moral concern. I further argue that this misclassification has practical consequences for AI safety, responsibility attribution, and policy design. By incorporating counterarguments from relational ethics, empirical studies of human–AI interaction, and global perspectives on moral status, the paper reframes AI safety as an ontological and ethical clarity problem rather than a matter of value encoding or moral optimization.